Horticultural tasks such as pruning and selective harvesting are labor intensive and horticultural staff are hard to find. Automating these tasks is challenging due to the semi-structured greenhouse workspaces, changing environmental conditions such as lighting, dense plant growth with many occlusions, and the need for gentle manipulation of non-rigid plant organs. In this work, we present the three-armed system HortiBot, with two arms for manipulation and a third arm as an articulated head for active perception using stereo cameras. Its perception system detects not only peppers, but also peduncles and stems in real time, and performs online data association to build a world model of pepper plants. Collision-aware online trajectory generation allows all three arms to safely track their respective targets for observation, grasping, and cutting. We integrated perception and manipulation to perform selective harvesting of peppers and evaluated the system in lab experiments. Using active perception coupled with end-effector force torque sensing for compliant manipulation, HortiBot achieves high success rates in our indoor pepper plant mock-up.

Haptic perception is highly important for immersive teleoperation of robots, especially for accomplishing manipulation tasks. We propose a low-cost haptic sensing and rendering system, which is capable of detecting and displaying surface roughness. As the robot fingertip moves across a surface of interest, two microphones capture sound coupled directly through the fingertip and through the air, respectively. A learning-based detector system analyzes the data in real time and gives roughness estimates with both high temporal resolution and low latency. Finally, an audio-based vibrational actuator displays the result to the human operator. We demonstrate the effectiveness of our system through lab experiments and our winning entry in the ANA Avatar XPRIZE competition finals, where briefly trained judges solved a roughness-based selection task even without additional vision feedback. We publish our dataset used for training and evaluation together with our trained models to enable reproducibility of results.

VR Facial Animation is necessary in applications requiring clear view of the face, even though a VR headset is worn. In our case, we aim to animate the face of an operator who is controlling our robotic avatar system. We propose a real-time capable pipeline with very fast adaptation for specific operators. In a quick enrollment step, we capture a sequence of source images from the operator without the VR headset which contain all the important operator-specific appearance information. During inference, we then use the operator keypoint information extracted from a mouth camera and two eye cameras to estimate the target expression and head pose, to which we map the appearance of a source still image. In order to enhance the mouth expression accuracy, we dynamically select an auxiliary expression frame from the captured sequence. This selection is done by learning to transform the current mouth keypoints into the source camera space, where the alignment can be determined accurately. We, furthermore, demonstrate an eye tracking pipeline that can be trained in less than a minute, a time efficient way to train the whole pipeline given a dataset that includes only complete faces, show exemplary results generated by our method, and discuss performance at the ANA Avatar XPRIZE semifinals.

Robotic avatars promise immersive teleoperation with human-like manipulation and communication capabilities. We present such an avatar system, based on the key components of immersive 3D visualization and transparent force-feedback telemanipulation. Our avatar robot features an anthropomorphic bimanual arm configuration with dexterous hands. The remote human operator drives the arms and fingers through an exoskeleton-based operator station, which provides force feedback both at the wrist and for each finger. The robot torso is mounted on a holonomic base, providing locomotion capability in typical indoor scenarios, controlled using a 3D rudder device. Finally, the robot features a 6D movable head with stereo cameras, which stream images to a VR HMD worn by the operator. Movement latency is hidden using spherical rendering. The head also carries a telepresence screen displaying a synthesized image of the operator with facial animation, which enables direct interaction with remote persons. We evaluate our system successfully both in a user study with untrained operators as well as a longer and more complex integrated mission. We discuss lessons learned from the trials and possible improvements.

Robotic picking from cluttered bins is a demanding task, for which Amazon Robotics holds challenges. The 2017 Amazon Robotics Challenge (ARC) required stowing items into a storage system, picking specific items, and packing them into boxes. In this paper, we describe the entry of team NimbRo Picking. Our deep object perception pipeline can be quickly and efficiently adapted to new items using a custom turntable capture system and transfer learning. It produces high-quality item segments, on which grasp poses are found. A planning component coordinates manipulation actions between two robot arms, minimizing execution time. The system has been demonstrated successfully at ARC, where our team reached second places in both the picking task and the final stow-and-pick task. We also evaluate individual components.

Part handling in warehouse automation is challenging if a large variety of items must be accommodated and items are stored in unordered piles. To foster research in this domain, Amazon holds picking challenges. We present our system which achieved second and third place in the Amazon Picking Challenge 2016 tasks. The challenge required participants to pick a list of items from a shelf or to stow items into the shelf. Using two deep-learning approaches for object detection and semantic segmentation and one item model registration method, our system localizes the requested item. Manipulation occurs using suction on points determined heuristically or from 6D item model registration. Parametrized motion primitives are chained to generate motions. We present a full-system evaluation during the APC 2016 and component-level evaluations of the perception system on an annotated dataset.

Locomotion in uneven terrain is important for a wide range of robotic applications, including Search&Rescue operations. Our mobile manipulation robot Momaro features a unique locomotion design consisting of four legs ending in pairs of steerable wheels, allowing the robot to omnidirectionally drive on sufficiently even terrain, step over obstacles, and also to overcome height differences by climbing. We demonstrate the feasibility and usefulness of this design on the example of the DARPA Robotics Challenge, where our team NimbRo Rescue solved seven out of eight tasks in only 34 minutes. We also introduce a method for semi-autonomous execution of weight-shifting and stepping actions based on a 2D heightmap generated from 3D laser data.

The predictive power of the h‐index has been shown to depend on citations to rather old publications. This has raised doubts about its usefulness for predicting future scientific achievements. Here, I investigate a variant that considers only recent publications and is therefore more useful in academic hiring processes and for the allocation of research resources. It is simply defined in analogy to the usual h‐index, but takes into account only publications from recent years, and it can easily be determined from the ISI Web of Knowledge.

In this paper a complete safety controller on a single chip is presented. HICore 1 is a comprehensive solution that includes a certified application specific integrated circuit for safety-critical applications according to the safety standard IEC 61508, meeting the safety integrity level SIL3. It also meets the requirements of the standard EN 13849 Performance Level e. A SIL3 operating system and a SIL3 middleware complement the presented safety chip solution. Based on the presented solution, the smallest certified safety controller represents an innovative product and allows system manufacturers to create safe solutions ready for certification.

The recently proposed fractional scoring scheme is used to attribute publications to percentile rank classes. It is shown that in this way uncertainties and ambiguities in the evaluation of specific quantile values and percentile ranks do not occur. Using the fractional scoring the total score of all papers exactly reproduces the theoretical value.

Fractional scoring has been proposed to avoid inconsistencies in the attribution of publications to percentile rank classes. Uncertainties and ambiguities in the evaluation of percentile ranks can be demonstrated most easily with small data sets. But for larger data sets, an often large number of papers with the same citation count leads to the same uncertainties and ambiguities, which can be avoided by fractional scoring, demonstrated by four different empirical data sets with several thousand publications each, which are assigned to six percentile rank classes. Only by utilizing fractional scoring does, the total score of all papers exactly reproduce the theoretical value in each case.

One way of evaluating individual scientists is the determination of the number of highly cited publications, where the threshold is given by a large reference set. It is shown that this indicator behaves in a counterintuitive way, leading to inconsistencies in the ranking of different scientists.

Using a very small sample of 8 data sets it was recently shown by De Visscher (2011) that the g‐index is very close to the square root of the total number of citations. It was argued that there is no bibliometrically meaningful difference. Using another somewhat larger empirical sample of 26 data sets I show that the difference may be larger and I argue in favor of the g‐index.

A percentile‐based bibliometric indicator is an indicator that values publications based on their position within the citation distribution of their field. The most straightforward percentile‐based indicator is the proportion of frequently cited publications, for instance, the proportion of publications that belong to the top 10% most frequently cited of their field. Recently, more complex percentile‐based indicators have been proposed. A difficulty in the calculation of percentile‐based indicators is caused by the discrete nature of citation distributions combined with the presence of many publications with the same number of citations. We introduce an approach to calculating percentile‐based indicators that deals with this difficulty in a more satisfactory way than earlier approaches suggested in the literature. We show in a formal mathematical framework that our approach leads to indicators that do not suffer from biases in favor of or against particular fields of science.

It is shown that under certain circumstances in particular for small data sets, the recently proposed citation impact indicators I3(6PR) and R(6,k) behave inconsistently when additional papers or citations are taken into consideration. Three simple examples are presented, in which the indicators fluctuate strongly and the ranking of scientists in the evaluated group is sometimes completely mixed up by minor changes in the database. The erratic behavior is traced to the specific way in which weights are attributed to the six percentile rank classes, specifically for the tied papers. For 100 percentile rank classes, the effects will be less serious. For the six classes, it is demonstrated that a different way of assigning weights avoids these problems, although the nonlinearity of the weights for the different percentile rank classes can still lead to (much less frequent) changes in the ranking. This behavior is not undesired because it can be used to correct for differences in citation behavior in different fields. Remaining deviations from the theoretical value R(6,k) = 1.91 can be avoided by a new scoring rule: the fractional scoring. Previously proposed consistency criteria are amended by another property of strict independence at which a performance indicator should aim.

Conduct-by-Wire is a vehicle guidance paradigm, which investigates the possibility of controlling an automobile by maneuver commands. The focus of this paper is to show one possible interaction strategy for maneuver-based vehicle guidance between driver and vehicle by means of discrete maneuvers. Therefore, the paper starts with a short introduction to the advantages of maneuver-based vehicle guidance and proceeds to the current design option for maneuver-based driver-vehicle interaction chosen by the TU Darmstadt. This includes the presentation of the user interface as well as the basic assumptions for maneuver-based interaction. This paper is finished by a specific example for the maneuver “Lane Change right”, which includes system behavior, as well as, information displaying and some restricting facts which are presented together with possible countermeasures.

A quickly increasing number of advanced driver assistance systems are implemented in modern vehicles. Already today, this can lead to a mental overload of drivers, since every single system has its own control elements. In the future, innovative vehicle interaction concepts, with a global design philosophy, will be necessary. Conduct-by-Wire (CbW) is an approach to relieve drivers from stabilization tasks, as the vehicle will be controlled by assigning concrete elements, the maneuvers, which are executed automatically. This could possibly lead to a more cognitive way of driving, keeping the driver closer in the loop. The basic requirement for this concept is the development of a maneuver catalog from the driver's point of view, in order to enable easy, even intuitive driving. Therefore, this paper describes how a driver would divide a specific track into single maneuvers, based upon an already existing catalog of commands. An innovative experimental concept is developed, consisting of a field and laboratory study, using the `thinking aloud' method, an eye tracking and a decision point analysis.

Abstract The g‐index is discussed in terms of the average number of citations of the publications in the g‐core, showing that it combines features of the h‐index and the A‐index in one number. For a visualization, data of 8 famous physicists are presented and analyzed. In comparison with the h‐index, the g‐index increases between 67% and 144%, on average by a factor of 2.

Wheeled tour guide robots have already been deployed in various museums or fairs worldwide. A key requirement for successful tour guide robots is to interact with people and to entertain them. Most of the previous tour guide robots, however, focused more on the involved navigation task than on natural interaction with humans. Humanoid robots, on the other hand, offer a great potential for investigating intuitive, multimodal interaction between humans and machines. In this paper, we present our mobile full-body humanoid tour guide robot Robotinho. We provide mechanical and electrical details and cover perception, the integration of multiple modalities for interaction, navigation control, and system integration aspects. The multimodal interaction capabilities of Robotinho have been designed and enhanced according to the questionnaires filled out by the people who interacted with the robot at previous public demonstrations. We present experiences we have made during experiments in which untrained users interacted with the robot.

Abstract J.E. Hirsch (2005) introduced the h‐index to quantify an individual's scientific research output by the largest number h of a scientist's papers, that received at least h citations. This so‐called Hirsch index can be easily modified to take multiple coauthorship into account by counting the papers fractionally according to (the inverse of) the number of authors. I have worked out 26 empirical cases of physicists to illustrate the effect of this modification. Although the correlation between the original and the modified Hirsch index is relatively strong, the arrangement of the datasets is significantly different depending on whether they are put into order according to the values of either the original or the modified index.

Abstract L. Egghe ( 2008 ) studied the h‐index (Hirsch index) and the g‐index, counting the authorship of cited articles in a fractional way. But his definition of the gF‐index for the case that the article count is fractionalized yielded values that were close to or even larger than the original g‐index. Here I propose an alternative definition by which the g‐index is modified in such a way that the resulting gm‐index is always smaller than the original g‐index. Based on the interpretation of the g‐index as the highest number of articles of a scientist that received on average g or more citations, in the specification of the new gm‐index the articles are counted fractionally not only for the rank but also for the average.

Abstract J.E. Hirsch (2005) introduced the h‐index to quantify an individual's scientific research output by the largest number h of a scientist's papers that received at least h citations. To take into account the highly skewed frequency distribution of citations, L. Egghe (2006a) proposed the g‐index as an improvement of the h‐index. I have worked out 26 practical cases of physicists from the Institute of Physics at Chemnitz University of Technology, and compare the h and g values in this study. It is demonstrated that the g‐index discriminates better between different citation patterns. This also can be achieved by evaluating B.H. Jin's (2006) A‐index, which reflects the average number of citations in the h‐core, and interpreting it in conjunction with the h‐index. h and A can be combined into the R‐index to measure the h‐core's citation intensity. I also have determined the A and R values for the 26 datasets. For a better comparison, I utilize interpolated indices. The correlations between the various indices as well as with the total number of papers and the highest citation counts are discussed. The largest Pearson correlation coefficient is found between g and R. Although the correlation between g and h is relatively strong, the arrangement of the datasets is significantly different depending on whether they are put into order according to the values of either h or g.

An essential capability for a robot designed to interact with humans is to show attention to the people in its surroundings. To enable a robot to involve multiple persons into interaction requires the maintenance of an accurate belief about the people in the environment. In this paper, we use a probabilistic technique to update the knowledge of the robot based on sensory input. In this way, the robot is able to reason about the uncertainty in its belief about people in the vicinity and is able to shift its attention between different persons. Even people who are not the primary conversational partners are included into the interaction. In practical experiments with a humanoid robot, we demonstrate the effectiveness of our approach.